104: Using AI Coworkers to Build Smarter Business Operations with Karl Simon

You may be late to the game and you'll never catch up. And let me ask you, and your opinion at that level, what are most companies getting wrong when they're starting to use AI at work? Well, I think singular task, singular activity, and that's missing the big impact. We work with a lot of clients. We go in. They're excited about using AI. Maybe some of them have started experimenting on their own, but there's usually not some orchestration occurring, right? How do they shift their thinking from that task level integration to workflow level orchestration? Yeah, usually it starts with calculating ROI. What's the difference in having people automate a singular task versus the orchestration of data to fulfill a business process that typically crosses not just systems, but the groups that own those systems? What is the approach to starting with systemizing, streamlining, whatever the orchestration is of these fragmented systems? The first thing to do is to stop buying another system to add to your frag stack. Because ultimately those tools, those fragmented systems, are likely to start disappearing one by one. Carl Simon is the co-founder of Subatomic Helping Companies, move beyond task-based AI into unified end-to-end workflows. He shows leaders how clean data AI co-workers and secure orchestration can eliminate business friction and unlock faster, smarter growth. Welcome to using AI at work. I'm your host, Chris Day. Each week we'll be learning how today's business owners, entrepreneurs, and ambitious professionals are getting more done with smart use of tomorrow's tech. Let's get started. Right now every business leader is asking the same question, what are we going to do about AI? If this is you, chiefayofficeer.com has the answer. We'll give you a simple path forward where we provide executive and team training so your people know exactly how to safely use generative AI in their day to day. We also manage the deployment and implementation to make sure tools actually get adopted and deliver results. And we'll also guide company-wide transformation, so AI becomes part of your operating system, not just another shiny object. The companies that act now will increase productivity, cut costs, and grow faster than their competitors. Those that wait, we'll get left behind. So if you want to make AI work in your business, visit chiefayofficeer.com and see how we're helping companies of all sizes finally get results from AI. Hey everybody, welcome to another episode of using AI at work. This is my favorite podcast, and I hope it's becoming yours as well. Today we are guest as Carl Simon, and guess what we're going to be talking about AI. And one of the things that I'm excited to dig in with Carl is how they are looking at AI holistically rather than task-based or activity-based. And they're really looking at ways to introduce AI into the entire workflow. Now some of the, I guess, challenges with that are going to be pretty universal, and that's the data environment. So anyway, Carl Simon is the co-founder at Subatomic, and that is what they do. Carl's belief is that AI should optimize end-to-end workflows across fragmented systems, not just improve those isolated tasks that probably a lot of us are working on. So Carl, before we get started, anything you want to let the audience know about what you're up to at Subatomic? Yeah, now I think you captured it perfectly, Chris. I mean, ultimately we're trying to help companies actually utilize the foundation that they have, which is data, but get it into a unified state, cleaned, standardized, so it becomes useful. When you introduce AI on top of poor data, you get poor results. You want good knowledge because those agents bring their intellect, and we at Subatomic bring those AI co-workers up to speed on your processes, your cognitive reasoning. All of it goes together. You have, unless you have strengths across the way from knowledge to intellect to your way, the difference heats you, you're not going to be positioned for success. So Carl, when you and I first did this pre-interview, I had a good idea of where this conversation was going. However, some recent developments have occurred that might change the direction of this episode. So I'm going to ask you something. And for those of you who are listening, if you've heard me talk about this before, please indulge me. I think this is a very important topic. Carl, are you familiar with what Jack Dorsey did at Block about a month and a half ago? You might need to summarize. Sure. For me. For the listeners, this is important. Again, indulge me. I might take a second. So Jack Dorsey Smart Guy started Twitter, started Block. A Block is the parent company of Cash App and Square. And there was an announcement on X. Block is laying off 4,000 of their 10,000 staff. To me, I thought, okay, AI got them, right? They introduced some efficiencies in certain areas and departments and they just don't need the people anymore. I wasn't that surprised. But that was February 26th. At the end of March, he released a paper on the website for Block, but also on the Sequoia website. Sequoia is a big investor in Block. And the title of the article was from hierarchy to intelligence. And in that, basically, he said, here's why we were able to fire those 4,000 people. He said, we did a deep dive on what, what does the organization of tomorrow look like? And he said, in our analysis, what we realized was that a big portion of this traditional pyramid-shaped hierarchy of boss at the top and then a layer under that, layer under that, layer under that was intended to collect information at a level and then pass that information up to the next level for evaluation, synthesis, decision-making, and then repeat that up and down the chain, up and down the chain. And he said, we realized in our organization, it was about 60% of the humans that we had, their job was moving information around. They might not have looked at it that way, but first principles indicated that that's exactly what they were doing. And he said, I had an aha moment. He said, this is what AI does. AI can ingest this information, evaluate, synthesize new outcomes or decisions to be made or whatever and then pass that to who needs it. But the challenge is the data. And he said, Block was uniquely situated to to go from this hierarchy into this what they're calling the intelligence layer because we were a remote company, every slack message, every email, every call transcript, everything that we were doing all day long was being captured as an artifact of our business, which is exactly the data that AI needs to be able to apply its intelligence contextually, accurately for your business. So the good news is Carl, this is like a boon for where subatomic is if more companies are going that way. And I'll tell you personally for chief AI officer, as soon as I heard that, I said, wait a minute, we were a remote company. We're doing that. And about five weeks ago, we started our version of preparing all of these business artifacts or being created to create this intelligence layer. Eric, see you out of Beverly Hills. He's a thought leader in the space. He calls it a single brain is the model. But our company chief AI officer is now doing this, meaning that our hierarchy of our org chart is about to flatten considerably. So if you were to wear that, know that that is and I talk to several people since. And I'm like, well, what do you think? And they're like, yep, that spot on. So Carl, let's jump into this. What do you think about that? Yeah, that's unsurprising to me. Hearing the reasons I wasn't up on that particular story, but I've heard many similar ones. And when you think about the old way we have done business that many are still doing today, it's the idea that all that information passing between people or up and down the hierarchy, yep, that's been built in friction all this time where it's not just a cost. There's just there are mistakes in transition and there are differences of the information that's being passed among groups that may not even align on business terms. The moment you have that unified data and it's all captured, everyone has visibility and accessibility to it. You don't need that those middlemen, those intermediaries. You know, it's you think about system integration. We talk about that a lot about sharing data. We have that the human level and that never made sense. But we never had a way to address it with technology before today. Yeah. So this sounds like a perfect environment for what you guys do at Subatomic, right? It does. I mean, we have always believed you got to get the data right and unified. Yeah. Set of terms, names, all. You know, Carl with the K with the C, Carla, I mean, my middle initial say. So some people get that messed up, right? I mean, who is Carl with the real Carl stand up? And, you know, having a unified language, a unified repository of information, whether it's structured or unstructured where everyone can actually retrieve and know it's saying the same thing, it's reflecting the same numbers means you don't have to present different information and wonder, did that person come to the meeting and give me the high level of how revenues being projected? know, where are key issues that are preventing us from closing more deals. Whatever the domain is in the use case that needs to be, you know, demonstrate through clear insights, no longer you have that issue. And the workflow of passing information, let alone executing the processes that make everything like revenue capture possible or operations fulfilled in the most efficient high velocity way. All those friction points disappear. If you got the right data is accessible, shareable. So sounds very attractive as a business owner. And for all of those who are listening to this right now, who are, you know, leading teams or leading entire companies, that sounds very interesting. Before and so at this point, I would say we've kind of covered a macro theme of this, the impact of having this data. Now let's bring it all the way back down to maybe the listener who is on the AI journey, but they certainly don't have any kind of like unified approach or there's still a lot of open loops or questions that they've got about this. Well, I think there are probably a few things I can, I can mention. You mentioned the beginning of the podcast, Chris, about singular task, singular activity. And that's missing the big impact. So they're getting that wrong. And so we recommend, you know, operationally, what's the most challenging thing you do that prevents you from actually reallocating your team members to more growth oriented activities. And then number two, the rollout itself, a lot of people think rolling out co-pilot, you know, and tracking usage is the way to go. But ultimately, people don't really know what to do. What the best practices are. And again, that fulfills more of the singular task aspect of create a piece of content for me or tell me about this spreadsheet. That's not really changing, you know, the, or moving the needle for your company to be operationally more efficient. And then finally, I think people think that just integration is sufficient instead of orchestration. I mean, in a unified data layer that brings it together with that singular view. When you try to just move data around to different systems in your existing, tool or tech stack, you're not really gaining the opportunity for that singular view where you can eliminate the friction. The friction's still there about people not understanding really the numbers correctly or in a very, you know, unified way. You use that term over and over again. And so I'll culminate or I'll summarize it with this key takeaway. The moment you have a unified view and can think about AI co-workers of fulfilling that unified, that unified view, let alone doing the work for you as you remain in control of the final output. You can finally actualize some material benefits and show that AI makes a big difference. Yeah. You know, this is it. What you said, the difference between integrating and orchestrating. We go in. Maybe some of them have, you know, started experimenting on their own, but there's usually not some orchestration occurring, right? They're seeing it at, they're evaluating it through this lens of the task level. Oh, there's these things that I could be using AI for. Oh, there's these things I could be using AI for. What does, what is required for those who are going to be leading this or kind of driving the message internally about what AI means? Yeah, usually it starts with putting an ROI, you know, calculating ROI. What's the difference in having people automate a single singular task versus the orchestration of data to fulfill a business process that typically crosses not just systems, but the groups that own those systems or functionally represent those systems. When you, when you finally get an organization that's crossing organization, you know, there are subgroup barriers. That's when true. You really can unlock the extended benefit sales of marketing across inclusive of service for customer service. You know, the aspects of manufacturing, inventory planning through manufacturing through the fulfillment process, you know, in the warehousing space. There are so many examples where you want to have that through line. We have visibility up and downstream of you and the workflows themselves are the connective tissue if the data is actually unified. So, you know, that's all ROI driven. You can find very quickly the opportunity to reduce the cost of the actual labor and execution. But even more importantly, over time, yet you realize that if you build it correctly, an AI comes to you where you work instead of you having to figure out how to work with AI. And what I mean by that is single pane of glass, you have access to all the information across those systems and perform, actually gets a focus with what you need to focus on for the week for today, before noon or even after like a chief of staff telling you, here's the key, here are the key things you need to work on, say, given tasks within the workflow, or, you know, clients you need to follow up with for whatever the reasons across compliance, service or revenue growth, then you can, you can enable that great unlock. So I just listened to an episode of a podcast by a guy named Nate B Jones. If those of you who are listening, don't know Nate B Jones, go to YouTube. I think you'll love his stuff if you like our podcast. But he talks about this concept of this proactive agent, right? Not and it sounds like that's what that's what we're going towards with this concept that you just introduced. So for those listening, who may not like, they maybe they get the concept or at least they understand the words, orchestrating fragmented systems. What does that look like in plain English? Well, I think what it means in plain English, if it's been built correctly in the back end, I get a little more into that in a moment. Then from a front end, from a usage perspective, you no longer are worried about where information is source from. You know that it's been built, unified standardized. And I feel like in some ways, I keep echoing what I've said earlier, but that's the key aspect of it where you can just actually start your day. You get a focus, chief of staff, we call it that or the concierge, we says, this is what you need to focus on. Or they are there to assist on any request that you give single pane of glass or font, you're, you'll find it in team, Slack, email, chat, team or text. However you want. They're there to serve you the way you like to work. You want to chase someone down the hall like a human being. You actually can actually, if you have subatomic on your phone, they're with you as you're running down the hall. Anyway, so you can interact with your AI coworker. Again, interesting. I mean, when people think of it as software, they think about it as having to go into a different application. When they think about it as an AI coworker, they realize they'll talk and communicate through all the existing communication channels they use today. That's a big difference in driving adoption and ease of use. And again, you hide all that complexity. Now, the back end, yeah, you want to actually land all your data raw from all those systems and then start the joining process, the cleansing process into a nice data like that is that standard cleansed, de-duped version of your data. And then create a dimensional model. Some people refer to it as the star schema data warehouse for the goal copy, which makes insight retrieval fast. And the operational, even for operation retrieval purposes for your workflow, instantaneous so that you can actually operate without friction. Again, you don't see it as a user. All that's been built for you. So it's ready to go. So, you know, when I think about going into a company and if they're the first thing they want to talk about is like, like organizing their data to me, that sounds like a really boring and heavy lift. What is required of a company that maybe they're using AI already, maybe it's fragmented, maybe it's not fragmented, but the, they're still thinking about it at the task level and not necessarily the workflow level. They're not thinking about it through this cross-system context that you're referencing here. What is the approach to, to starting with systemizing, streamizing, whatever the orchestration is of these fragmented systems? Yeah. The first thing to do is to stop buying another system to add to your frag stack because ultimately those tools, those fragmented systems, are likely to start disappearing one by one or at least you're going to reduce the licenses needed to go directly into those tools. But then the next step is sitting down with Subatomic and doing a two-week discovery. We'll talk about those workflows that are costly today and you describe exactly the way you do it today. Now, if you have great documentation on that, already fantastic. But if you need a document, the great thing about working with Subatomic is we'll auto-generate your as-is process today and then compare it to what it can be tomorrow. And that means that as long as we can get that written down and that will be consumed by our AI co-workers who do all the work behind the scenes to engineer and deliver solutions to you. And we identified where the data comes to fulfill those workflows already and think about that in terms of the back-end unified data later engineering. Nothing is really stopping you from moving forward today. You just have to identify which workflows to begin first. And again, stack ranks in terms of the ROI for you. Yeah, that's a great place to start. So, you know, one of the, I guess, pushbacks we get as interested as a company as excited as they are, one of their questions is, well, how much time is this going to take from my day? I'm already busy, right? And I think you and I, Carl, we know that, look, if you do the thing, you'll open up bandwidth just but you got it. You know, there's going to be a little pain initially. Now, you mentioned this in your case with Subatomic coming on site or working with the team, there's a two-week window. What happens in that two weeks? It's just interviewing everyone to get a better sense of how the workflows operate today, what data they rely on to fulfill the data workflows. And then, so, it's not two weeks where every single person will be in interview mode for eight hours per day. It's just a two-week period where we'll get the right stakeholders and the right performers, the SMEs, you know, in those organizations that can represent the process and form about what's done, give the tribal knowledge, which has probably never been documented, so we can capture that into the cognitive engine that will be built for you and your AI coworkers that execute for you. So, it's on a total per person or per team or the per discovery basis. It depends on the number of use cases we're tackling within that two-week period, but it's not a full-time. I mean, maybe you'll spend 20% of your week working with us in that discovery period. On-going post-discovery will be an iterative process with you week to week, and we can meet one sort twice per week, and it would probably be an hour max per meeting, maybe 30 minutes, for the right people who are going to be operationally keeping track of where we are on delivering, and to provide the feedback that we need to incorporate through each iteration. Because ultimately, we build it fast. Now, the data loading and unification, all that logic is spun up fast, but the load itself of data could take time, depending upon the amount of history we're talking about, but from week one after discovery, you already start seeing the solution with Subatomic. Our AI coworkers build out the actual pipelines from all your source systems into that unified data layer. They build up the workflow, we capture and incorporate the cognition that you use for reasoning and decisioning, while having captured already your SOPs for the 2B state that we're building to replace the current friction-filled existing processes. A company that's going to approach this, I think the takeaway is that you don't flip a switch. There is some work required. There is going to be participation from your subject matter experts, from the domain experts within your organization, if you want to do this right. If you don't want to invest that time, the best you can get is this drive-through window of getting people chat GBT and say, "You can go ahead and ask a question, get an answer, and then bye-bye." That is about as far as you're going to be able to get with AI, unless you do something like what Carl's talking about, which is this orchestration, this intelligent design being put into these efforts, not just the next tool. And you mentioned that frame gets that. Fundamentally, that's no different than the way it's always been, except now AI auto-generates the delivery of the solution and helps you iterate fast. So the opportunity to seize those material benefits actually compresses in terms of time. Yeah, but you know, I think that there's, especially for people who aren't in the AI industry, they're running their business and AI is, there's a little bit of FOMO and they know they need to do something about it. Maybe they're using it and getting some wins. There's this false narrative of, "Oh, AI will do it for me. Oh, AI is going to, but the reality is, until the co-pilots, the agents that are working with you, the co-workers, take over, the human still has to drive this stuff. So I want to make sure that anybody listening, that's just how the sausage gets made. If you want to do it right, you're going to have to participate with the initial heavy lift. But once that's done, if you've done it right, you get benefit immediately. If you don't do it and you delay it and you want to continue to get the new tool or try to figure out something and still have your AI processes primarily still human in the loop, you're never going to get past that. So my advice, and I think Carl would probably bite the bullet, invest those two weeks, the week, whatever it is for your company to do the heavy lift so that you can't get the benefit. One hundred percent. I know that you're not pitching speed at any cost. That's why there's this kind of two-week environment. One of the things that you also reference, which I like a lot, we tell clients, eat the elephant one bite at a time. You're not starting initially with doing this across the entire business. We're targeting the R.O.I. the likely R.O.I. environments where if we can speed this up, if we can compress the sales cycle, if we can have more communications with prospects and clients, then we'll generate more revenue. But outside of those, where are some of the places where people should expect to see the initial efforts be deployed? What departments? What activities? Well, we fulfill for any domain, any industry, any use case. But a lot of our initial traction has happened in the wealth management R.A.A. Space. And so what we look to do to help them prepare for their clients where they have many offerings, for example, financial management, tax management, estate planning, retirement planning, and so forth. There's a lot of work that goes into preparing for quarterly meetings with their clients, or understanding which clients actually should have that frequency for a direct advisor client relationship, or otherwise be placed into a service center for more frequent touches, but not necessarily in person meeting the meeting. But the collective amount of information that is required to make sure that it's aligned with what they've done in the past, what their goals are, what the latest actions were from the prior meetings, and a better understanding the macro and micro-economics and how it affects their specific portfolios. It takes at least four hours per client that's been our experience with our clients. And so there's an enormous number of hours that can be saved there, which means you can move towards growing your book instead of just preparing for the next meeting. So whether it's that or account opening, which crosses many systems, I think a great way to think about where the opportunities are, and I know I'm saying this in more of a general statement is when you have five to ten or even more systems out there where you have to touch these different systems today to do the work, you know you have an opportunity to unify your view of the data and get that true orchestrated firm going. Those are likely going to be the opportunities, but we see it across again, the operational prep, and I gave the example on the RIA space. But there's plenty of opportunity to streamline sales and marketing. I think today, even with the best SaaS players, it's still a disconnected experience between sales, marketing, service. And there's an opportunity to get that unified as well. In fact, at Subatomic, we didn't like any CRM that existed out there, so we built our own. And we're likely to commercialize that now for the market to also purchase, because it'll be more of that single view and you have your chief of growth officer that guides you and focuses where you need to focus. Tell you who to meet with not just from your least prospects and client perspective, independent of whether or not it's with a sales or marketing focus, but internally with your sales and marketing and service teammates. Finally, eliminating that friction. Yeah, you know, I guess it was about a year ago, we were sponsoring a visted. event in Chicago. And I had no idea how much wealth management, you know, businesses or firms were in Chicago. And we ended up talking to a multinational bank that was there looking to grow their assets under management. And one of the things that we do for clients is we say, "Okay, we're going to, we call it a McKenzie, a mini-McKenzie. We kind of run on a McKenzie-style analysis on their business, particularly their industry, to look for some low-hanging fruit across primarily sales, general, and administrative activities, S-GNA. And, you know, we had recently been doing a lot with construction companies and that sort of thing. And we could move the needle on primarily looking at what's the impact going to be on EBITDA, right? And for a construction company on the high end, it was like maybe 18 to 20% lift on EBITDA. Pretty good. But the more average was maybe 8 to 12%. Still, better than a sharp stick in the eye. But when we ran this for wealth management, it was a 35 to like 60% lift on EBITDA because, I mean, it's obvious now, the bulk of the expenses associated with that industry are the people. And it's their time. These are highly compensated individuals. Like 80% of what they spend is the people. It's not necessarily a technology or, you know, equipment or leases. It's the time of these highly compensated individuals. And if AI can step in, like what you're talking about right there, like I know for a fact, we saw that that the impact it could have and registered investment advisory or wealth management, any financial services, huge. So I think that you guys are targeting the right thing. Now, we, I don't know. I see a trend really with what's preventing companies from being more aggressive about their AI implementation and deployment. The first one is, I love to use it, but I'm not quite sure where. And it sounds like you're, you're too weak. Discovery is the identification of the where. The second thing is, I love to use it, but I don't understand the risk fully. Right. So if I don't understand the risk, especially with the industries that you're targeting, better safe than sorry, right. It's easier to say no and not expose ourselves to the risk. So what does it mean to actually build AI with compliance and security as, you know, because everybody's, everybody's interested in the, the efficiencies and the, the ROI impact and the ebit of lift, like I talk about. And it's like an almost an afterthought with the compliance and security side of things. So what does that look like to build that in from the beginning? Yeah, I can tell you what it looks like. It's subatomic. I think you're right, Chris. A lot of people think about it secondarily. And we treat those both compliance and security as first class citizens as subatomic. So a lot of companies not named subatomic will say we log and trace everything, but you don't really have visibility to it. So in other words, it's there in the event of discovering something after the fact, you can actually use as raw data and try to pull up some analysis. They don't give you visibility. They don't give you key insights. They don't give you proactive and preventative correction that would be applied to compliance or full prevention defense ability when it comes to security as subatomic. That's exactly what we do. Every single request you make is actually traced dynamically. And then auto generated give you full visibility to the step wise things that happen behind the scenes, including for a given step, the full reasoning trace on what occurred and the decision made relative to the different options that could have been chosen given that scenario. That level of trace ability is not just good for our clients to give that visibility and get comfortable. It's important to us as subatomic and our AI co-workers that will course correct on the fly before it gives you that final result. Not to mention at the request level for compliance and security for that matter, you can see all the security checks that were performed for that given request. We track all those individual transactions at an overall aggregate level, full dashboarding into security and compliance. Dashboards within dashboards drilling into the actual data and the executions so that you can actually see everything that happened within a single pane and our insights analyzer, which is another AI subatomic co-worker that comes apart of your foundational core subscription will identify opportunities to course correct and flag things that require fulfillment in terms of compliance or added security. I think it's a very sound approach here. Now, you mentioned the concept of AI co-workers a lot and anybody listening to this has heard the term agents and agentic AI and those sorts of things. But what do you mean by an AI co-worker? What does that mean in subatomic lexicon? Yeah. Why? You can think of agents and AI co-workers as synonymous and instead of AI co-workers making a full agentic team for the agentic workflow. But we really distinguish our AI co-workers because they come, you know, first of all, immersed in your SOPs and your reasoning, your cognition on how you like to make decisions and perform and differentiate. But secondarily, because they've been trained on how to apply and perform with compliance and security in mind, they are, you know, subatomic AI co-workers are very different than your typical agent off the shelf that you can get somewhere else. And for those trying to build it themselves, that's awesome. But to get it with compliance and security foundation that's a tough thing to do. I can concur. Anybody who's tried to build an open claw and have it break every five minutes understands this. So I like the language of an AI co-worker. An agent sounds very like I picture that robot, you know, emoji, right? But an AI co-worker, oh, you know, that sounds a lot less threatening. But how should somebody who's managing teams that include AI co-workers, how should they think differently about their role when they've got, you know, AI co-workers instead of, you know, 50 butts and seats? Right. They should think of themselves as now at an managerial level. I mean, let's roll back the tape to the beginning of the discussion, Chris, where you highlighted dorses, you know, reasoning of why AI actually reduced employees, right? All those management layers disappear because you didn't need all that friction happening. But the thing is, and the reality is, it's not just giving information to everyone in unified singular state that everyone sees that is a benefit. It's actually getting all the mundane work done for you. But now you as an individual contributor of the past need to be the manager. You need to be the domain expert that knows whether or not your AI co-workers are performing correctly. Because you are, you must be the human in the loop. You must provide feedback so your AI co-workers learn along the way. And so encoding, I mentioned this earlier, Chris, we have AI co-workers engineering everything. Now, we have engineers on our team, but I look to them as the key architects and the key managers of the AI co-worker personnel that actually generate the solutions for them. They need to make sure that they understand the design patterns that should be employed because it's best for this particular unique solution and the architecture that's inherently a part of the end to end. So you read, you level up, you know, your performance using all the things you've understood in the past and done yourself. But now you're managing AI co-workers to do it for you so you can get everything done faster. Okay. Well, then in that case, as a the human, what skills are becoming more valuable when I'm managing AI assisted workflows, AI co-workers? So being able to understand and recognize when an AI delivery or output actually matches your expectations, what would you have looked for in your output? And it should be really aligned again at the group level or corporate vision so there are standards within any organization that filter down of what's expected from quality accuracy, the compliance that's required everything. You need to be really thoughtful now of always be the verifier now of how your AI co-workers are working and, you know, actually the coach, you know, with the feedback you provide, not just for course correction, but to improve the way you like to have the outputs done because you know your human manager, your direct, you know, direct manager actually prefers the information provided and presented in a very specific way. How has it been for the individual sub-atomic comes in there introducing this stuff? How has it, has it been an easy transition? for the humans to be able to move into that role of the coach for the AI co-workers? You know, nothing is ever 100% of the time perfectly easy, but most of them have, again, for a couple of reasons, number one, they're not changing the way they work, necessarily, in terms of working in teams. Got it. It's like, e-mail or a single pane of glass simplifies the world. And when that last piece of that last consideration, where whether it's in teams or whether it's within the subatomic single pane of glass, again, their time gets reallocated to actually just coaching up their AI co-workers, which is not a big effort, by the way. I mean, if they weren't already applying best practices in the role, they were unlikely producing good outputs, which means they have to be taught what the organization overall wants. But as long as they've always applied it, then it should be automatic to be able to make those same checks you would do on your own work, but now done on AI co-worker, by AI co-workers. Technically, it's still your work. Yeah. It's still accountable and responsible for the output. Nothing has changed, other than you've gotten a lot of help getting stuff done. And this is going to be important, not just for today, but increasingly as we move through the months, quarters and years, just the velocity of change is going to be critical as companies are actually accelerating timelines to deliver. Think about the way the early days of mobile phones might have had a new phone every two years and it became every one yet. I mean, you're going to have compressed schedules of delivering, whether you're in products, whether you're in marketing for messaging and positioning. It's going to constantly be evolving fast. Again, I said this, did I say this in the beginning? Subatomic helps you adapt evolve and scale. We decided to build ourselves, hire people who can adapt evolve and scale, build solutions to do that because we want our clients to. That's the future and the future is coming fast because if your competitors are already adapting evolve and scaling, you've got some cash rep to do. You need to start running at that speed. Yeah, but let me ask you, can you catch up? If your competitors are already doing this, and I mean, I've thought about this, can you catch up? If you're like, oh, we'll start next quarter. We'll start, you know, Q4, whatever. But you've got competitors, even if it's messy right now, they're going to figure this out. And when they do, their velocity of production and just increase in bandwidth, like, can you catch up? That will be a very important thing to watch. I'm predicting it'll be very hard. The later you get started, the faster that there's sooner that others have already adopted AI and you begin adapting evolving, you may be late to the game and you'll never catch up because they'll their velocity will be iterating, you know, multiple numbers ahead and you're just trying to catch up. That's, I talk about it in terms of imagine, this is the gap between your competitor who's ahead of you and this is where you are. It's going to, even when you start trying to catch up, the gap is going to continue to widen because they're at it. Yes. Yes. Much more advanced mature level of true efficiency, true growth. And they're able to absorb these, like, the new releases that come out faster because they've already created the infrastructure and the culture and the training of their people to where they're used to, oh, there's an update. We're now going to do, you know, this is faster. We don't have to do that one step anymore. So, and one thing that I think that is important to note that if you're listening to this and you're like, well, the good news is in my industry, people aren't really moving fast. You don't know. And I'll tell you what I mean by that. Your competition has called somebody like me. And we've been on site for, you know, the kickoff and we've been working with their teams weekly for maybe it's been a month. Maybe it's been two or three months. And they're not out there getting billboards saying, hey, we're practicing with AI. Hey, we're using AI. You, this is all happening unintentionally in stealth mode. Your competition is doing this stuff. So if you think that, you know, you've got time, I'm telling your brother, you don't. It's time to like, it's time for you to like bang the table and say, guys, like, if we're not doing something now, we need to do something. Let them listen to this and understand what's possible, right? So, you know, this, this strategic agility that's going to be coming with this stuff. Where are the companies struggling to change these like what's, once there, there's this opportunity to introduce this AI-driven workflow. Where is the friction that a, that a business owner should expect while this becomes the new way of doing things? I think number one is everyone's fearful for their job. And again, this is the opportunity for business leaders to help create a culture of acceptance that AI is coming and then support that your leaders actually want you to thrive in the world of AI. And so, starting to think about how do I manage AI co-workers? That's number one. Number two, I would say AI leaders need to reduce the change management aspect of rolling out a new solution. Don't choose new solutions that require your human workers to actually learn something new. Again, AI should work for you. You should not work for software of any kind like we have in the past. Yeah. Now it's the time for AI to meet you in Microsoft Teams Slack email, right? Simplify their world adoption becomes easier. And then finally, you know, having, you know, a self-service source of information always learned about how AI can help you. I think if you put that educational learning system that is again, you know, self-service, but also perhaps mandatory at certain scheduled moments or milestones in your time at the company, you minimum need to be up skill to a certain level. Then I think that will create the foundation where people are less fearful with the aspect of, okay, I know I have to do this, but how do I get started? Right? How do I keep up and learn and contribute to the rapid velocity our organization is likely to undergo because we're starting to use AI and and reap the benefits? Change is a foot, no doubt. Well, Carl, thank you so much for sharing what you guys are doing at Subatomic. How can individuals who are recognizing that their data situation is not ready for this and they need some help how can they find out more about what you guys are doing at Subatomic? Yeah, so there are many ways. I mean, one of the key ways is go to getsubatomic.ai. That's our website. You can go to LinkedIn, find me or anyone in Subatomic. If you look for the company Subatomic, you should find our company and and the employees there. But whether it's Carl Simon on LinkedIn, I'm Silva. He's the other co-founder will be available to talk to you and assess your top opportunities for AI to actually materialize benefits sooner than later for you. Then finally, you can actually email me a Carl, that's with the K, Carl with the K, at getsubatomic.ai. So I'm on your website and I love this headline. Everyone's buying AI, smart firms are hiring it. I like that a lot. Yeah. And so for all the listeners, we're going to have the links to all this information in the show notes. And I just want to say thank you again for having this podcast be part of your AI upskilling. And if you're getting value out of this and you know somebody else, co-worker, a peer, a friend, a golf buddy, whatever, please share this with them. This is how we get the opportunity to help other people expand their perspectives on what's possible with AI. So just want to say thank you so much for being a listener. And if you have a second, leave us a review forward an episode along and make sure that you join us for the next one. Carl, thank you so much for being here and everybody will see you on the next episode. Thank you Chris. Thanks for tuning into using AI at work. Don't forget to subscribe for more conversations about how to use AI at work. And a special thank you to our sponsor Chief AI Officer for empowering businesses with AI education and training. Visit their website for a free AI readiness assessment and AI strategy guide to help you get started using AI at work. That's www.chiefaiofficeer.com. Follow us on Twitter and the handle using AI at work. And visit www.usingaiawork.com for free resources to help you harness AI in your role.